RXNorm - SNOMED Alignment
Please provide comments on the examples below representing the challenges in harmonizing the US Drug Extension with SNOMED CT international release introduced during the 2025-02-20 DEUSG meeting. Complete details can be found at 2025-02-20: DEUSG Virtual. Slides can be found at: https://drive.google.com/file/d/1rkpoLnmKMecO1Bi5xGJVteT2miJy2vZ0/view. Recording starts at 55:00.
Summary of 5 examples:
Thiothixene: SNOMED has tiotixene hydrochloride but RxNorm has Thiothixene (as active ingredient and BoSS). Hydrochloride form is no longer in US market.
Capmatinib: RxNorm has capmatinib hydrochloride but SNOMED has capmatinib (as active ingredient and BoSS).
Quinacrine/mepacrine: RxNorm uses the US preferred term of mepacrine. Also, SNOMED specifies the salt/hydrochloride, but RxNorm has the substance as active ingredient and BoSS [same as #1 issue]. Hydrochloride form is not approved in US market.
SNOMED is using 108890005 |Botulinum toxin type A (substance)| to model clinical drugs, while RxNorm is using 703368006 |IncobotulinumtoxinA (substance)|. Are these two substances duplicates? RxNorm uses the United States Adopted Name (USAN).
Odevixibat: RxNorm uses a hydrated form, odevbixibat sesquihydrate, while SNOMED uses odevixibat (as active ingredient and BoSS). Also, RxNorm has oral pellet form while SNOMED has oral capsule.
Comments
Hi Nicholas,
In relation to reference sources for adding drug content, we always look for official sources of drug information by licencing/drug regulatory bodies in the first instance. One of the sources of information we would look for, for products licensed in the US, would be the FDA - approved drug labelling information (which seems to be comparable to the summary of product characteristics (SPC) published by the regulatory bodies EMA and MHRA in EU/UK and Product Information (PI) published by the Therapeutic Goods Administration in Australia.
If there are other official sources you use please include. We need information to include the active ingredients and the BOSS to allow us to model medicinal products in line with our editorial policy but also other information such therapeutic class and mechanism of action etc to ensure correct modelling of the corresponding substance and subsequent placement in the hierarchies.
I have looked at the 5 examples provided where alignment is proving difficult and there does not seem to be an easy fix on first glance. Some of the differences arise due to differences in editorial policies in expressing salts and hydrates and also in the case of Odevixibat - discrepancies in expressing dose form. Product information (SPC) below stipulates dose form as hard capsules (and the capsules contain pellets).
https://www.ema.europa.eu/en/documents/product-information/bylvay-epar-product-information_en.pdf
As Matt has alluded to, there may be discrepancies between product information and difficulty in reviewing legacy drugs that are no longer available in many countries such as Thiothixene and unlicensed drugs in use such as Mepacrine hydrochloride in the UK.
Botulinum toxin type A products need further analysis to ensure correct representation. The FDA approved information actually stipulates Botulinum toxin type A as the active ingredient for both DYSPORT (but labelled DYSPORT® (abobotulinumtoxinA) and XEOMIN (incobotulinumtoxinA).
In a nutshell, the five examples have highlighted several issues causing difficulties in providing a solution for alignment - the drugs are no longer marketed, differences in editorial policy and inconsistencies in product information.
I would also welcome discussion how we proceed.
tl;dr - I think this comes down to:
inconsistency/reliability in product documentation
terminologist debates about what is a clinically significant salt vs base (I use salt in the very loose sense of compound/variant)
limits to modelling the legacy SNOMED CT content.
Thiothixene, Capmatinib, Quinacrine,Odevixibat
We don't appear to have any of these drugs in Australia. However, we've encountered similar problems for other drugs. And I think it's the product of a combination of a few things:
International content being too specific for some products - precise ingredients at the Clinical Drug level (rather than Real Clinical Drug)
Debate over what's a clinically significant salt, and what should be at the MP level.
Fully modelling legacy content based on the FSN in the absence of real world product documentation.
The unreliability of real world product documentation.
Even within the International content it can be confusing:
785079005|Betamethasone valerate 1 mg/mL cutaneous cream|
785080008|Betamethasone valerate 1 mg/mL cutaneous lotion|
374372000|Betamethasone (as betamethasone valerate) 1 mg/g cutaneous cream|
374014000|Betamethasone (as betamethasone valerate) 1 mg/g cutaneous lotion|
Also note the different units. We also encounter this when different manufacturers use different units. So it's not simply a cream vs lotion issue.
These products have different BoSS.
In AMT we only have BoSS = Betamethasone for these products.
Our preferred strength representation (for topicals) is 0.1% for these. Fortunately all our Clinical Drugs are "mg/g" other wise we could potentially end up with duplicate PTs. (or need to specify %w/w)
Then there's the question of what the "active ingredient" is. Clinical significant salts/compounds.
INT has a single MP 778502002|Betamethasone only product| subsuming a range of precise formulations
Betamethasone dipropionate
Betamethasone sodium phosphate
Betamethasone valerate
Whereas AMT has determined these compounds to be clinically significant and created three specific MPs accordingly.
Wether this remains correct, I'm not sure. But a lot of debate often goes into determining if a salt is clinically significant.
Some of that though comes down to what is the purpose of concepts at that general "Medicinal Product" level?
(Perhaps AMT's approach is too specific? And needs to be reviewed)
Botulinum toxin - This is an interesting (bot not novel) issue, and I don't think they're duplicates.
The TGA describes 5 substances that have a synonym "Botulinum toxin type A" (all with different CAS No.)
abobotulinumtoxin A
incobotulinumtoxinA
onabotulinumtoxinA
prabotulinumtoxinA
relabotulinumtoxinA
Yet AMT describes two different products
Xeomin (incobotulinumtoxinA 50 mouse LD50 units) powder for injection, 50 mouse LD50 units vial
Botox (botulinum toxin type A 50 mouse LD50 units) powder for injection, 50 mouse LD50 units vial
The ingredients described are as per the documentation in their registration.
I can find Botox packaging that mentions onabotulinumtoxinA but I can also find boxes that don't.
And again, this comes down to reliability/availability of documentation and decisions about what is clinically significant.
I don't think their duplicates. But I'm not sure if their siblings or parent/child... that requires more research.
Thanks @Matt Cordell and @Karen Rees for this feedback. I have relayed this information to our RxNorm team for review. Will share their feedback ASAP but I will not have a response to share for tomorrow's DEUSG meeting as many of our RxNorm SMEs are out of office this week.
The US NRC/NLM requests comments and feedback from DEUSG members and the SI content team on the possibilities aligning these 5 example clinical drugs based on the input given by the NLM RXNORM team. As part of the comments and discussion, we hope to learn how SI International drug content is added from their sources and would be nice to know what those sources are. Thanks in advance for your help!